A mechanistic understanding of cancer rests heavily upon the mutated genes. These include genes causing familial tumor susceptibilities, governing replication fidelity, and residing in regulatory pathways. Pancreatic cancer is an unusually efficient system in which to identify distinctive mutational targets, discoveries to which our research group has contributed heavily. This success is due in part to highly informative structural patterns of homozygous deletions as well as a higher incidence of genome-maintenance mutations (especially those affecting homologous recombination) in this tumor type. Indications are that many recessive genes and perhaps additional dominant genes remain to be discovered in pancreatic cancer. Most of the high-risk pancreatic cancer families remain unexplained by known mutations, and most cancers having CIN (chromosomal instability) are not yet tied mechanistically to a molecular cause. Recessive mutations in novel genes, for example, are currently being discovered more rapidly than can be annotated for their functional significance. We recently explored and published key technical breakthroughs and special tumor cell resources that are newly available and can quickly accelerate this line of study. Powerful high-throughput sequencing techniques are underway to aid such studies, but will need to be complemented by a roadmap derived from a strategic structural analysis of the cancer genome such as we are developing. Our specific aims will locate promising sites of new mutant genes. For genes having some existing functional clues, we can assess the effects of mutation using available assays. For truly novel genes having few clues as to their function, we can achieve missense mutations or gene disruption using technologies we developed for homologous recombination in somatic cells, with phenotypic assessment accomplished both by gene-specific assays and by orthotopic xenografting. For genes having unknown functional assignment and distant evolutionary conservation, a zebrafish model of gene impairment will be used to survey for developmental clues to function and for an ability to augment zebrafish pancreatic tumorigenesis. This latter technique seems surprising, but appears to offer immense efficiency for classifying novel genes. Our long-term goal is to provide a more complete foundation for future studies of familial susceptibility, disrupted signaling pathways, and genome instability in pancreatic cancer.